NewSparseWANDIndex()

This function creates a SPARSE_WAND index configuration for sparse vector search using the Weak-AND algorithm.

func NewSparseWANDIndex(metricType MetricType, dropRatio float64) Index

PARAMETERS:

• metricType (MetricType)

  The distance metric type for similarity search (e.g., index.COSINE, index.L2, index.IP).

• dropRatio (float64)

  The ratio of small vector values to drop during indexing. Range: [0, 1).

RETURNS:

Index

An index configuration instance. Pass this to CreateIndex() via the index option.

Example

import (
    "github.com/milvus-io/milvus/client/v2/index"
    "github.com/milvus-io/milvus/client/v2/milvusclient"
)

// Create index configuration
idx := index.NewSparseWANDIndex(index.COSINE, 0.2)

// Use with CreateIndex
createIdxOption := milvusclient.NewCreateIndexOption("collection_name", "vector_field", idx)
task, err := client.CreateIndex(ctx, createIdxOption)